CineSat´s Outstanding Analysis & Forecast Products
		Atmospheric Motion Fields Predicted Satellite Images Predicted Trajectories Cloud Contour Prediction Cloud Development Maps Convective Cells Satellite Image Movies
	Atmospheric Motion Fields
		CineSat´s forecasts are based on motion fields that describe the movement of clouds from one image to the next one. They are derived from successive half-hourly or hourly images of a geostationary meteorological satellite. The major processing steps involved in CineSat´s cloud motion analysis are Cloud pattern tracking Automatic quality control of motion vectors, and Iterative improvement of meteorological motion field quality
	Real-Time Cloud Pattern Tracking
		For a number of cloud segments in the current satellite image, CineSat searches in a local neighbourhood of the previous image to find the position of the most similar piece of cloud. This intermediate product are raw cloud (or atmospheric) motion vectors.
			Accurate, fast, validated Quality controlled vectors Good MSG performance
		The major achievement in this area is a pattern tracking method that is about 20 times faster than the commonly used correlation techniques. This makes it possible to achieve processing times for thousands of wind vectors within only several minutes - even on small workstations or PCs.
	Motion Field Quality Control
		Pattern tracking provides perfect results in case of non-overlapping, non-developing objects that follow a rigid motion. But in case of natural scenes, additional methods have to be applied to take into account the floating nature of cloud structures. In particular, when new clouds appear and others disappear, a simple pattern tracker does not find a correct motion vector. The CineSat approach is to perform a quality control and quality improvement of the raw cloud motion vectors. The following tests are applied to each single motion vector: Local consistency checks Timely consistency checks Timely consistency of cloud pattern content Significance of cloud texture Significance of the located position All test values, e.g. the local consistency quality value, are derived only from neighbouring (or predecessor) vectors representing similar temperatures. They are then mapped to a total quality value by a test specific arctan -shaped quality function. Each single test gives a quality indicator between 0 and 100%. The individual tests are finally combined by a weighted sum to a final quality indicator. The single quality test weights are determined according to the intended further use of the vectors: For Cloud Motion Winds the tests are trimmed to identify those vectors that most probably represent a real wind. For nowcasting, a consistent multi-height description of cloud motion is required, thus preferring vectors at the borders of cloud structures. This difference between wind and cloud movement is important, since cloud motion is related to but not strictly determined by winds -consider e.g. a strong wind blowing over a mountain chain, but the clouds standing still at the mountain ridge for a long time. Meteorologists report that CineSat perfectly handles such a situation whereas a wind based forecast would fail completely.
	Cloud motion versus wind at a mountain ridge: Wind speed in red Cloud motion in green
		For nowcasting you need to accurately estimate the position of clouds and cloud systems in the near future - and exactly this information is provided by CineSat´s cloud motion forecasts.
	Motion Field Quality Improvement
		This is the most important step when bridging from raw cloud motion winds to nowcasting products - bad pattern matching vectors have to be replaced by meteorologically better ones.
		Depending on user parameters, CineSat iteratively replaces all vectors below a given quality threshold by vectors with better quality with respect to the selected quality criteria. This recursion of pattern matching and quality control is quite computing intensive, but results in a meteorologically consistent vector field - allowing for multi-level motion, and being quite different from simply smoothing out a raw cloud motion wind field.
	Cloud Motion, System Motion, and Relative Motion
		CineSat let's you compute three types of motion fileds, i.e. AMF - the absolute motion of small clouds SMF - the movement of larger cloud systems, and RMF - the movement of clouds within cloud systems
		AMF = SMF + AMF
	Top	There are several options available to define, detect and confine cloud systems. You can use large cloud patterns to track the system motion or compute the mean vector of all clouds within a given mask or from all neighboring clouds that have similar properties Having the absolute motion AMF and system motion SMF you can compute the relative motion either as difference of both fields (RMF = AMF - RMF) or by compensating the system motion and computing the residual motion The large choice of methods reflects the current research status of this feature family. They are provided to encourage user's of the CineSat system to take part in this research and to select the optimum method for their practical applications.
	Predicted Satellite Images
	Top	One of the most unique features of CineSat is the prediction of satellite images several hours into the future. Predicted cloud images are an intuitive way to efficiently convey nowcasting information to a forecaster. Looking at an animation loop over the most recent satellite images and extending this animation some hours into the future gives the forecaster an instant and intuitive understanding of the expected flow of clouds and upper tropospheric air masses. CineSat Nowcast does this extrapolation of the most recent cloud movements in very much the same way a human interpreter would go - but based on a more in-depth analysis of the cloud motion field, i.e. 1) Trend analysis of the last few hours of cloud motion 2) Prediction of future cloud motion 3) Application of predicted cloud motion fields to image data Although quite good results are obtained when basing the prediction on the last motion field only, CineSat now allows to use the history of the last hours of motion fields - usually the last 2 hours. Motion information older than 2 hours seems not to contribute significantly to prediction accuracy. The algorithm that moves image pixels according to a predicted motion field has been specifically tailored to the meteorological image content. Studies at meteorological sites have shown that CineSat´s forecasted images are convincing and meteorologically plausible for up to: Meteosat Image Channel Operational Prediction Visible channel 1.5 - 2 hours Infra-red channel 2.5 - 3 hours Water vapor channel 2.5 - 4 hours	Predicted Images in user defined time steps unique CineSat feature efficiently convey nowcasting information intuitive understanding of air mass flow for the next 2 to 4 hours Have a look into the future
	Predicted Trajectories
		CineSat´s trajectories describe the expected future path of clouds for the next few hours. In addition to the predicted images, trajectories are an important means for visualizing cloud motion. Their benefit is that the movement can be shown in a static single picture. They are especially useful for communicating nowcasting results in printed form. In principle, CineSat takes the start points set by the user and moves these points according to the predicted motion fields several hours into the future. Predicted trajectories can be computed at user defined scales and positions, but can also be applied to synoptic measurements (e.g. rainfall), to lightning positions, weather radar cloud positions, and image features like convective cells.	Trajectories expected path for the next few hours in user defined time steps can be applied to user-defined starting points
	Two-hour predicted trajectories in four half-hourly steps on a regular grid over Sicily, Italy. Top
	Cloud Contour Prediction
		CineSat extracts the contour lines in the current IR image at user defined temperature thresholds and computes the 2-hourly predicted contour lines by use of the cloud motion field. This nowcast product is displayed as an overlay on the current IR image.
		This way, a single display gives you a fairly good overview about expected cloud coverage and the dynamic cloud movement during the next hours. Of course, you can adapt the default forecast range of 2 hours to your own requirements.
	Cloud Development Maps
		Meteorologically relevant changes in cloud images are a combination of cloud motion and cloud development. CineSat´s image change detection tool separates cloud movement from cloud development. Cloud development includes the emergence, enlargement, shrinking, deformation, temperature change and disappearing of clouds. Cloud Development Maps can be computed for user defined time intervals, i.e. usually in the range of 30 minutes up to 2 hours. With CineSat, for the first time ever, meteorologists can immediately separate cloud movement from cloud development:   Cloud movement is shown by the CineSat motion field   Cloud development is presented by the CDM image       Having the cloud development map available, you can immediately focus your attention on areas of interesting on-going developments without the need to scan the entire image. You will get immediate pointers to upcoming convective cells. An example is given in the figure below.	Image Change Detection separates cloud movement from cloud development user defined intervals precice information about on-going cloud development
	Cloud development map of exploding thunderclouds over Africa   The overlay shows: new developing clouds in red and disappearing clouds in blue   The blue and red zones indicate real changes of cloud systems that are not due to cloud movement. Top
	Convective Cells
		CineSat performs a detailed automatic analysis of convective cloud cells using the infra-red image channel. In addition to the CB position, you get a description of the cell size, its temperature, development stage, moving direction and speed, and the expected further path. The result data are available as user configurable overlay, as well as a table containing the list of cells and cell properties.
	Predicted Trajectories of convective cells (and CBs - Cumulonimbus) over Europe.   The overlay shows: path of high (cold) CBs in red path of medium cells in yellow path of low (warm) cells in green   The CineSat trajectories predict the 2 hour cell movement. Top		CB Analysis automated CB detection size & position temperature development trend direction, speed, future path
	Satellite Image Movies
		Having two satellite images - at half-hourly or hourly distance - and the motion field in between them, it is an easy exercise to interpolate the movement of the single pixels down to e.g. 1 minute steps. Thus, CineSat makes the single clouds realistically moving and developing along their correct path from one real image to the next. Conventional animation techniques build image loops by simply fading the half-hourly satellite images. This results in a flickering animation in all regions where clouds move. CineSat overcomes this problem by computing the intermediate images based on the cloud movement information. You get sharp and smooth films with realistically moving and developing clouds. Broadcasting companies confirmed that the products obtained with this technology are significantly better and far more attractive than those obtained by standard techniques.	Perfect Image Movies 1-minute steps no flickering realistically moving & developing clouds sharp & smooth films best quality for your media clients
	real 12:00 UTC Meteosat WV image interpolated 12:30 UTC image real 13:00 UTC Meteosat WV image
		This sequence demonstrates how a small cloud in front of the Portuguese coast moves and develops from the first image over the interpolation to the last image.
	Top
	Overview Benefits Air Traffic Control Applications Forecast Products Features FAQ Whats New ? Config / Install The Server Distributors Contact Us
	Order your free CineSat Demo CD Copyright © GEPARD, J. Scheiber KG 1988 - 2010